Effects of Flow and Bulk Vesicle Concentration on Supported Lipid

Subscriber access provided by Gothenburg University Library

Article

Effects of Flow and Bulk Vesicle Concentration on Supported Lipid Bilayer Formation Christina M. Bailey, Anubhav Tripathi, and Anita Shukla Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b02764 • Publication Date (Web): 26 Sep 2017 Downloaded from http://pubs.acs.org on September 27, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Effects of Flow and Bulk Vesicle Concentration on Supported Lipid Bilayer Formation Christina M. Bailey, Anubhav Tripathi, Ph.D., Anita Shukla, Ph.D.* School of Engineering, Center for Biomedical Engineering, Institute for Molecular and Nanoscale Innovation, Brown University, Providence RI Abstract Supported lipid bilayers (SLBs) have been used extensively in a variety of biotechnology applications and fundamental studies exploring lipid behavior. Despite their widespread use, various physicochemical parameters have yet to be thoroughly investigated for their impact on SLB formation. In this work, we have studied the importance of flow in inducing the rupture of surface adsorbed chicken egg derived L-α-phosphatidylcholine (egg PC) vesicles on silica and gold surfaces via quartz crystal microbalance with dissipation monitoring (QCM-D). On silica at 25°C, egg PC vesicles were found to adsorb in a flattened configuration (~13 nm thick compared to bulk vesicle diameters of ~165 nm) but only undergo a transition to a stable SLB under flow conditions. In the absence of flow, an increase in system temperature to 37°C was able to promote vesicle rupture and SLB formation on silica with a 10 times lower rupture time compared to rupture under continuous flow (175 µL/min flow rate). Gold surfaces, with their increased hydrophobicity, led to less vesicle flattening once adsorbed (~60 nm thick structures), and did not support vesicle rupture or SLB formation even at flow rates of up to 650 µL/min. We also showed that under continuous flow conditions, vesicle adsorption rates on silica surfaces follow Langmuir kinetics, with an inverse dependence on bulk vesicle concentration, while an empirical power law dependence of vesicle rupture time on bulk vesicle concentration was observed. Ultimately, this work elicits fundamental insight into the importance of flow and bulk

ACS Paragon Plus Environment

1

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 49

vesicle concentration in the adsorbed vesicle rupture process during SLB formation using QCMD.

*Corresponding author: [email protected], Tel. (401) 863-5719

Keywords: L-α-phosphatidylcholine vesicles; supported lipid bilayer; quartz crystal microbalance with dissipation monitoring; flow; bulk vesicle concentration


2

Page 3 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Introduction Supported lipid bilayers (SLBs) can serve as models of the phospholipid cell membrane commonly used in biotechnology applications.1–3 SLBs have been used to study molecular interactions of lipid bilayers with various macromolecules and small molecules.4–7 Previous work has investigated SLB formation with varying lipid compositions,8 incorporation of lipid proteins in SLBs,9–11 and utilizing polymer cushions12 to yield SLBs that mimic specific aspects of cell membrane function. The great potential of lipid bilayers for biomedical applications has been demonstrated with investigations spanning molecular and nanoparticle transport,4,5 protein assembly and function,9–11 peptide arrangement and insertion,6,7 drug screening,13 and microfluidic platforms.14 In order to optimally utilize SLBs in these and other advanced applications, it is necessary to fully understand the bilayer self-assembly process. Previous studies have investigated the formation of SLBs on various surfaces such as silica,1 mica,15 gold,16 and titanium.17 Properties of the substrate material can influence whether the vesicles remain adsorbed to the surface or rupture, leading to eventual formation of a planar bilayer structure. For example, silica and mica are often used to induce vesicle rupture, while more hydrophobic surfaces, such as gold, have been utilized to maintain intact adsorbed vesicles.18 Additionally, studies have examined the impact of varying lipid compositions,19 osmotic effects,20 the presence of divalent ions,21,22, gel-liquid domain effects,23 and lipid transition temperatures2 on vesicle adsorption and rupture abilities. While these studies have led to several significant advances in the SLB field, more work needs to be done to enhance understanding of factors contributing to the different stages of vesicle rupture and SLB formation at a fundamental level.


3

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 49

Ensemble techniques, such as quartz crystal microbalance with dissipation monitoring (QCM-D), which is used extensively in the formation and monitoring of SLBs, typically indicate vesicle rupture once a critical vesicle coverage (CVC) has been reached.1 Using techniques such as atomic force microscopy, isolated rupture of single vesicles into bilayer islands has been observed.24,25 The role of vesicle fusion and coalescence has also been discussed, where two or more neighboring vesicles merge in order to develop a complete SLB.26,27 The goal of the current study was to investigate the role of flow, temperature, and substrate properties in the SLB formation process, along with examining the dependence of vesicle adsorption and rupture times on bulk vesicle concentration. Here we utilized QCM-D to precisely control physical and chemical parameters during phospholipid vesicle adsorption to provide further insight into the driving forces governing SLB formation. QCM-D measures frequency and dissipation changes occurring in real-time on a substrate consisting of a coated piezoelectric quartz sensor. This is achieved by applying a voltage over the QCM-D substrate inducing resonance; frequency and dissipation are monitored at odd intervals of the fundamental frequency.26 Briefly, to monitor the vesicle adsorption to SLB formation process via QCM-D, vesicles are administered to an enclosed QCM-D module (Figure 1a), where they are allowed to interact with the substrate and the change in frequency and dissipation are continually monitored. In this study, we investigate the importance of flow in promoting QCM-D SLB formation. We demonstrate that in the absence of flow, increasing thermal energy can promote SLB formation. To the best of our knowledge, the specific effects of flow on SLB formation in a QCM-D system have not yet been reported. By using a range of lipid concentrations, we also investigated the impact of bulk vesicle concentration on vesicle adsorption and rupture kinetics.


4

Page 5 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Materials and Methods Lipid vesicle formation, vesicle characterization, and surface characterization Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich. MilliQ water (18.2 MΩ, EMD Millipore, Taunton, MA) was used in all experiments. L-α-phosphatidylcholine (Egg, Chicken) (egg PC) (Avanti Polar Lipids, Inc., Alabaster, AL) in chloroform was stored at 20°C and used to form lipid vesicles, as previously reported.28,29 Briefly, stock egg PC was dried under ultra-high purity nitrogen gas (Corp Brothers, Inc., Providence, RI) in a chemical fume hood and kept under vacuum for at least 4 hours to remove excess chloroform. Dried egg PC was rehydrated in Tris sodium chloride (NaCl) buffer containing 10 mM Trizma base and 100 mM NaCl, yielding a final concentration of 2.5 mg/mL (3.2 mM) egg PC in Tris NaCl (pH 7.8). The rehydrated egg PC was subsequently subjected to 5 freeze-thaw cycles, freezing at -78°C with dry ice and thawing at 45°C on a heat block with vortexing between each cycle, yielding multilamellar vesicles (MLVs). To form large unilamellar vesicles (LUVs), this egg PC MLV suspension was extruded 21 times (Avanti Polar Lipids Mini Extruder, Alabaster, AL) through a 100 nm pore size polycarbonate membrane (Nuclepore Track-Etch Membrane Filtration Products, Whatman, Maidstone, United Kingdom). LUVs were stored under nitrogen gas at 4°C and used within one week of extrusion. Prior to experimental use, egg PC LUVs were diluted to the desired concentration using Tris NaCl. Dynamic light scattering (DLS) using a Malvern Zetasizer ZS90 with Zetasizer 7.01 software was used to examine the hydrodynamic size and diffusivity of the MLVs and LUVs in Tris NaCl at 25°C at concentrations of 0.01, 0.1, and 1 mg/mL. Cryo-transmission electron microscopy (TEM) was also used to examine the egg PC LUVs for 1 mg/mL samples. Vesicles


5

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 49

were deposited on a specially prepared lacey carbon electron microscopy grid, transferred under liquid nitrogen to the cooled tip of a cryo-transfer stage (CT3500J; Oxford Instruments, Concord, MA), and imaged using a JEOL 1200 TEM as previously reported.30 A goniometer (Ramé-Hart, Succasunna, NJ) was used to measure the static contact angle of the egg PC LUVs (0.1 mg/mL in Tris NaCl), Tris NaCl, and water on silica-coated and goldcoated substrates (Biolin Scientific, Västra Frölunda, Sweden). Silica surfaces were cleaned with a water, 2% w/v sodium dodecyl sulfate, and water rinse sequence, followed by drying with N2 and UV/ozone treatment with a UV/ozone ProCleaner (Bioforce Nanosciences, Salt Lake City, UT) prior to contact angle measurements. Gold surfaces were cleaned with a 3:1 (v/v) ammonium hydroxide: hydrogen peroxide mixture (i.e., TL1 cleaning), rinsed with MilliQ water, dried with N2 and UV/ozone treated prior to contact angle measurements. DROPimage software (Ramé-Hart, Succasunna, NJ) was used to analyze the static contact angle after deposition of 2 µL of liquid on the surface. Quartz crystal microbalance with dissipation monitoring (QCM-D) measurements A QCM-D E4 system (Biolin Scientific, Västra Frölunda, Sweden) was used to monitor in situ lipid vesicle adsorption and bilayer formation by measuring frequency changes (∆F) and dissipation changes (∆D) occurring on quartz QCM-D substrates over time. Figure 1 provides a schematic of the QCM-D chamber geometry and shows how the QCM-D substrate interacts with vesicles introduced into the QCM-D chamber. Changes in frequency correspond to the oscillation of the piezoelectric quartz crystal (with fundamental frequency of 5 MHz) sandwiched between gold electrodes over which an alternating voltage is applied. Silica- and gold-coated quartz crystals (Biolin Scientific) were utilized in this work. Prior to use in QCM-D experiments, the substrates were cleaned in the same way as they were cleaned for contact angle


6

Page 7 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

measurements (see section Lipid vesicle formation, vesicle characterization, and surface characterization). Change in mass adsorbed on the crystal, including bound water, can be obtained using Sauerbrey31,32 or Voigt33,34 modeling analysis of these frequency changes, depending on whether the adsorbed film is more rigid or viscoelastic, respectively. Dissipation changes are measured when the applied voltage is turned off and the dampening of the acoustic waves is monitored over time. This energy dissipation is related to the viscoelasticity of the sensor-adsorbed material. For all vesicle adsorption and bilayer formation studies, a baseline frequency and dissipation measurement was first established in Tris NaCl for a period of 10 min. Egg PC LUVs were then introduced into the QCM-D flow chamber and the frequency and dissipation measurements were continued. QCM-D measurements under flow conditions QCM-D monitoring of egg PC LUV behavior under flow conditions was carried out at 25°C on both gold- and silica-coated QCM-D substrates. For experiments with gold substrates, a single egg PC concentration (co) of 0.1 mg/mL in Tris NaCl was examined at a pre-adsorption flow rate (Qo) of 175 µL/min. Following vesicle adsorption, Tris NaCl was flowed through the system at an increased flow rate (Q) of 650 µL/min. For experiments utilizing silica-coated substrates, six separate egg PC LUV co values of 0.01, 0.0167, 0.05, 0.1, 0.3 and 0.5 mg/mL, with a constant flow rate (Qo and Q) of 175 µL/min, were examined. QCM-D measurements without flow The impact of flow on vesicle behavior was further analyzed by introducing a no flow experimental condition. No flow QCM-D studies were conducted at two temperatures, 25°C and 37°C, using silica-coated substrates with a co of 0.1 mg/mL and Qo of 175 µL/min. For the


7

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 49

elevated temperature studies, all solutions were heated to 37°C in a heating block prior to flow through the QCM-D modules. Vesicle flow was stopped once a small amount (∆F ~ -4 Hz) of egg PC LUVs were introduced into the QCM-D chamber (not corresponding to complete substrate coverage) and ∆F and ∆D were continuously monitored. An additional delayed flow introduction condition was examined at 25°C, in which the QCM-D system was allowed to incubate for a period of at least 15 hours once flow was stopped, following which, flow was introduced with Tris NaCl flowing at a Q of 175 µL/min. ∆F and ∆D were continuously monitored. Statistical analysis and mathematical modeling All experiments were conducted in triplicate at minimum. Representative results for the 3rd overtone are shown in figures for all QCM-D ∆F and ∆D measurements unless otherwise specified. Results are given as mean ± standard deviation and statistical significance was calculated using one-way analysis of variance (ANOVA; α = 0.05 and p < 0.05 was considered statistically significant). QTools 3 software (Biolin Scientific, Västra Frölunda, Sweden) was used to determine the mass and thickness of the adsorbed egg PC structures at different time points using the ∆F and ∆D QCM-D data gathered. A custom MATLAB (MathWorks, Natick, MA) algorithm was also used to further analyze the transient raw frequency and dissipation data obtained from QCM-D measurements.

Results and Discussion Characterization of egg PC lipid vesicles and substrate surface properties Lipid vesicles composed of egg PC were developed using the dry lipid thin film technique.28 Prior to utilizing these vesicles in QCM-D studies, the successful formation of vesicles was characterized using cryo-TEM imaging and DLS. Cryo-TEM imaging confirmed


8

Page 9 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

the unilamellar structure of the extruded LUVs (Figure 2a) and suggested a vesicle diameter of approximately 100 to 150 nm. In order to confirm the decrease in the hydrodynamic diameter of egg PC LUVs compared to egg PC MLVs, DLS was performed on both the MLVs obtained prior to extrusion and the LUVs obtained after extrusion (Figure 2b). DLS demonstrated a z-average of d = 165 nm ± 10 nm and d = 2552 nm ± 1617 nm for LUVs and MLVs, respectively. The average standard deviation of the size distribution within one sample based on an average of three measurements was ± 57 nm and ± 345 nm for LUVs and MLVs, respectively. The LUV polydispersity index (PDI) was 0.08 ± 0.03, while the MLV PDI was 0.31 ± 0.11. No significant difference in the z-average size range was observed for DLS conducted at different vesicle concentrations (data not shown). The diffusivity, Dv, of egg PC LUVs in Tris NaCl reported by the DLS was found to be 2.94 × 10-8 cm2/s. We also estimated this diffusivity using the Stokes-Einstein equation for a sphere,

=

(1)

Here, kB is the Boltzmann constant, T is the temperature, η is the viscosity, and d is the vesicle diameter. Using Equation (1), for T = 25°C, η estimated as the viscosity of water at 25°C (η = 8.90 × 10-4 Pa ⋅ s), and d = 165 nm, we obtained a D of approximately 2.64 × 10-8 cm2/s, which is in close agreement with the Dv reported by the DLS. To better understand vesicle adsorption and rupture on the QCM-D substrates of interest, we also investigated the wetting properties of silica- and gold-coated QCM-D substrates. The static contact angle of water, Tris NaCl buffer, and egg PC LUVs on these substrates was measured and is shown in Table 1. As expected, the silica substrates were significantly more hydrophilic than gold in each of these test conditions (p < 0.0001 using one-way ANOVA with α


9

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 49

= 0.05). The addition of egg PC LUVs to Tris NaCl did not appreciably alter wettability of the substrates. The water contact angle on silica was not measurable due to the rapid spread of water on the silica substrate. Table 1: Contact Angles on QCM-D Substrates. QCM-D substrate coating Silica

Gold

Solution Water Tris NaCl buffer Egg PC LUVs (0.1 mg/mL) Water Tris NaCl buffer Egg PC LUVs (0.1 mg/mL)

Contact angle (°°) Not measurable 13.4 ± 2 17 ± 4 74 ± 5 64 ± 4 64 ± 4

Effect of bulk vesicle concentration on adsorption and rupture of lipid vesicles Using QCM-D, an egg PC lipid bilayer was observed to form under continuous flow of Qo = 175 µL/min on silica-coated quartz substrates. We investigated the impact of varying bulk vesicle concentration, co, from 0.01, 0.0167, 0.05, 0.1, 0.3 and 0.5 mg/mL on both the vesicle adsorption and rupture processes, key to eventual bilayer formation.1,15 Figure 3a shows both the QCM-D frequency change (∆F) during the vesicle introduction to bilayer formation process as well as dissipation change (∆D) for odd overtones (n), 3, 5, 7, 9, and 11 for a co of 0.1 mg/mL. Frequency changes relate to mass and thickness of the adsorbed material.34 Dissipation relates to the viscoelastic properties of the adsorbed material,35 and is defined as:

=

(2)

where El is the energy lost during an oscillation cycle and Es is the total stored energy. Figure 3a shows that at co = 0.1 mg/mL, lipid vesicles adsorb without rupturing on the silica surface until they reach a critical vesicle coverage (CVC) at ∆F = -45 ± 4 Hz for n = 3. CVC has previously been described as the amount of vesicles on a surface required to trigger


10

Page 11 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

vesicle rupture and the start to bilayer formation.36,37 During this process, ∆D increases from 0 to 5.3 × 10-6 ± 0.4 × 10-6 for n = 3 due to adsorption of the softer vesicles on the rigid silica substrate. After CVC is achieved over an adsorption time period of ta = 4.5 ± 1 min, the ∆F increases to -26 ± 1 Hz over a 3 ± 1 min time period (for n = 3), which we will refer to as the rupture time period, tr. During this time, ∆D decreases to 0.4 × 10-6 ± 0.2 × 10-6, indicating vesicle rupture into a SLB, a process accompanied by the release of entrapped water and excess lipids.38,39 The bilayer is significantly more rigid than the non-ruptured vesicles and this low dissipation value is indicative of the bilayer being closely coupled to the substrate. The decrease in separation of ∆F traces over time for the different overtones following the rupture period also confirms the increased rigidity of the SLB as compared to the adsorbed vesicles.40 Figure 3b and 3c show a comparison of the ΔF and ΔD QCM-D data obtained for co = 0.01, 0.1, and 0.5 mg/mL on silica-coated quartz substrates, for n = 3. ΔF and ΔD QCM-D data obtained for all concentrations examined (co = 0.01, 0.0167, 0.05, 0.1, 0.3, and 0.5 mg/mL) on silica-coated quartz substrates, for n = 3 can be found in Figure S1. Compared to a ta of 4.5 ± 1 min for co of 0.1 mg/mL, a co of 0.01 mg/mL, 0.0167, 0.3 and 0.5 mg/mL led to ta of 17 ± 1 min, 12 ± 1 min, 1 ± 0.2 min, and 0.9 ± 0.4 min, respectively. These adsorption times are much greater than the diffusion times of these vesicles to the surface, given by Equation (3). ~ 2

/

/

(3)

Here, Dv was obtained from the DLS (2.94 × 10-8 cm2/s), Mw is the molecular weight of egg PC (768 g/mol), NA is Avagadro’s number, and Nagg is the average number of lipid molecules per LUV, which was estimated to be 2.7 × 105 using an egg PC lipid molecule area of ~0.64 nm2.41


11

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 49

With these values, we estimate td ~ 1.5 s. Thus, vesicle adsorption onto the surface is the ratelimiting step in the overall process leading to SLB formation. Although the adsorption of vesicles onto a surface is a complex process, we find that Langmuir kinetics with a negligible vesicle desorption rate can accurately describe the relationship between adsorption time, ta, and bulk vesicle concentration, co. Assuming Langmuir kinetics, the vesicle adsorption rate is proportional to the number of vacant adsorption sites (Γ∞ -

Γ) and the bulk vesicle concentration, co, in the buffer. Here Γ∞ is the total concentration of vesicle adsorption sites on the surface [mg/cm2] and Γ is the surface concentration of vesiclebound sites [mg/cm2]. Hence, the rate of vesicle adsorption can be written as:

=

! "# (Γ&

− Γ)

(4)

Here, k1 is the vesicle adsorption rate constant. We have assumed co is in excess compared to the adsorbed vesicle concentration. Equation (4) can be integrated to obtain: ) =

!

*

ln (

-

- .

)

(5)

Equation (5) suggests that ta is inversely proportional to co. Figure 3d confirms that indeed the adsorption time is inversely proportional to the bulk vesicle concentration in the concentration range examined in this work (R2 = 0.9751). Rupture times, tr, have been shown to depend on vesicle size, initial vesicle geometry, and the chosen material substrate.15,16,18 We found that tr also depends on the vesicle concentration when flow rate and surface properties are fixed, where tr = 8 ± 1, 8 ± 2, 2.7 ± 0.5, 3 ± 1, 1 ± 0.1, and 0.8 ± 0.3 min, for co = 0.01, 0.0167, 0.05, 0.1, 0.3, and 0.5 mg/mL, respectively. An empirical correlation of tr versus co is shown in Figure 3e, indicating a power law relationship between tr and co (R2 = 0.9489): / = 0.528"# .4.5 !

(6)


12

Page 13 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Due to the rigidity of the SLB, we can determine the change in adsorbed mass (∆mf) on the oscillating surface using a normalized Sauerbrey Equation:35,42 ∆78 = −9∆:

(7)

Here, C = 17.8 ng cm-2 Hz-1, which is the Sauerbrey mass sensitivity constant arising from the quartz crystal fundamental frequency (5 MHz). From Equation (7), using the final ∆F data point obtained upon SLB formation, we determined that mass of the SLB formulated for all bulk vesicle concentrations examined is 460 ± 14 ng/cm2. As expected, there is no dependence of SLB mass adsorbed on co, as all bulk vesicle concentration values lead to a similar SLB ∆F, despite the dependence of the adsorption and rupture times on co. This mass includes the hydration layer sandwiched between the silica-coated QCM-D sensor and the SLB. For SLBs formulated using 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), this hydration layer was previously reported to have an areal mass of ~102 ng/cm2.43 Using this value as an estimate for the mass of the hydration layer of the egg PC SLBs formulated in this work, we estimated the areal mass of the egg PC SLB without the hydration layer as ~358 ng/cm2. We used these adsorbed masses to obtain an estimate of the SLB thickness with and without the hydration layer, using Equation (8): ∆=>

;>

Here, deff is the estimated SLB thickness and ρeff is the effective film density. We used the same

ρeff utilized by Rodahl et al. (800

A

=B

) in their analysis of lipid vesicle adsorption for egg PC

vesicles.44 Using Equation (8), we estimated the thickness of the SLB to be ~5.8 nm and ~4.5 nm with and without the hydration layer, respectively. Thus, the entrapped water is ~1.3 nm in thickness, similar to what has previously been reported.45,46 The egg PC SLB bilayer thickness


13

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 49

we estimated is in agreement with previous reports for SLBs formulated from 4:1 (v/v) DOPC:1,2-dioleoyl-sn-glycero-3-phospho-L-serine (DOPS) (~4.1 nm thickness)1 and 1palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) (3.5 nm thickness).47 Effect of flow on lipid vesicle rupture To elicit mechanistic insight into the significance of flow in the vesicle adsorption and rupture process, we performed QCM-D experiments at 25°C with a co of 0.1 mg/mL using silicacoated quartz substrates, in which flow was stopped following the introduction of vesicles into the QCM-D chamber (corresponding to ∆F~-4 Hz). Figure 4a shows the frequency and dissipation changes for odd overtones, n = 3, 5, 7, 9, and 11. The vesicles were incubated within the QCM-D chamber over a period of at least 15 hours in the absence of flow, following which flow was introduced at Q = 175 µL/min. During the no flow incubation, vesicles adsorbed on the surface reaching a plateau ∆F = -55 ± 2 Hz for n = 3 at approximately 10 hours. From Figure 3a, in a continuous flow system we have seen that once a CVC corresponding to a ∆F of -45 ± 4 Hz is obtained (for n = 3), vesicle rupture follows, leading to SLB formation. However, without flow, we observe the vesicles to remain on the surface as a monolayer of non-ruptured vesicles even after exceeding the CVC noted for the continuous flow system and reaching the plateau ∆F. The dissipation changes occurring in Figure 4a are similar to the continuous flow system where we observed an increase in dissipation during the vesicle adsorption phase, here reaching ∆D = 8 × 10-6 ± 1 (for n = 3). When flow is introduced following the no-flow incubation period, a rupture of the vesicles occurs leading to SLB formation, as indicated by the rise in ∆F to approximately -26 ± 1 Hz and decrease in ∆D to 0.4 × 10-6 ± 0.2 × 10-6 for n = 3 in Figure 4a. Figure 4b provides an overlay of ∆F and ∆D data obtained during the continuous flow experiments (i.e, the same


14

Page 15 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

experimental conditions whose representative data is shown in Figure 3a) and the delayed flow introduction experiments (i.e., the same experimental conditions shown in Figure 4a) following the commencement of the rupture process. The rupture process begins once the CVC is reached for the continuous flow condition and once flow is introduced after the no-flow incubation period for the delayed flow introduction experiments (set to time = 0 in Figure 4b). From Figure 4b, we see that the rupture to SLB formation process takes the same amount of time in both the stop flow and continuous flow conditions and occurs at similar rates. Estimating adsorbed vesicle layer thickness The dynamics of the viscoelastic adsorbed vesicle layer prior to rupture, with a thickness, ha(t), can be modeled using the Voigt model to fit the QCM-D ∆F and ∆D output.33,42 The important properties of the adsorbed vesicles and the semi-infinite bulk Newtonian liquid utilized in this modeling shown in Figure 1b are the density, ρa, shear viscosity,ηa, and elastic modulus,

µa, of the adsorbed vesicles, and the density, ρf, and viscosity, ηf, of the bulk fluid. The observed change in frequency and dissipation energy due to the presence of the viscoelastic film can be derived using the Voigt model as follows:33,44 C=(D)

∆: = ?

,

E E

(9)

and F X>

W = O

X .O> X>

and J8 = UY

G?> >

Page 16 of 49

(14 and 15)

Note, ρq is the quartz density (2648 kg/m3) and tq is the thickness of the quartz substrate (0.33 mm). The QCM-D sensor oscillates at resonant frequency of ω = 2πf0 = 10π MHz. The Newtonian liquid properties are based on water properties at 25°C as ηf = 8.9 × 10-4 Pa⋅s and ρf = 1000 kg/m3. The sensitivity of ha(t) on ρa, ηa and µa of the adsorbed vesicle layer was examined using a custom MATLAB algorithm utilizing Equations (9-15). Figure 5a and 5b show changes in ∆F and ∆D versus ha, for various ρa, in the range of 800 to 1200 kg/m3. As expected, ha depends strongly on ρa. However, ∆D is relatively insensitive to changes in ρa. Figure 5c and 5d show changes in ∆F and ∆D versus ha for various µa, in the range of 4 × 103 to 4 × 106 N/m2. We see that ha is most sensitive to changes in µa in the range 4 × 104 N/m2 to 4 × 105 N/m2. ∆D, however, is a strong function of elastic modulus and changes non-linearly with ha over the entire range of µa examined. Figure 5e and 5f show changes in ∆F and ∆D versus ha for various ηa in the range of 1 ×10-3 to 1.05 × 10-2 Pa⋅s. ∆F and ha are almost insensitive to changes in ηa. The energy dissipation of the adsorbed layer decreases with its viscosity. The experimental ∆F(t) and ∆D(t) data obtained during the no-flow incubation period in the delayed flow introduction experiments, in which vesicles are adsorbing to the QCM-D substrate, were fit using the Voigt model. Figure 6a shows the comparison between theoretical and experimental ∆F and ∆D using the best-fit values of ρa = 800 kg/m3, ηa = 0.007 Pa⋅s, and µa = 4 × 104 N/m2. Figure 6b shows the Voigt model estimate of ha versus time during the no-flow incubation. The plateau thickness obtained from Voigt modeling for the adsorbed vesicle layer during the no-flow incubation was 13 ± 3 nm. Compared to the original diameter of the vesicles


16

Page 17 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(~ 165 nm), this thickness value suggests that the vesicles are arranged in flattened pancake-like structures on the surface. Lind et al. previously suggested that a structure consisting of vesicles adsorbed onto an already formed SLB may not be distinguishable from an adsorbed layer of only vesicles based on QCM-D ∆F and ∆D data.2 In our studies, if this SLB plus vesicle structure was a major contributor to the CVC layer prior to the start of flow, then flow would serve to simply remove the additional vesicles, revealing the characteristic SLB ∆F and ∆D traces. However, these structures are prominent when experiments are conducted below the lipid melting temperature (Tm). Egg PC has a Tm below 0 °C,48,49 and all experiments in this work were performed well above this temperature. Thus, it is unlikely that these structures are contributing significantly to our pre-flow surface. Thermodynamics of egg PC SLB formation It has previously been shown that flattened adsorbed lipid vesicles may either fuse with other flattened vesicles or remain as single flattened vesicles on the surface. These vesicle structures can eventually rupture forming single-bilayer patches that can coalesce and lead to SLB formation.24,50 Energetically, the bound vesicles will rupture if the free energy associated with adsorbed vesicles, Fvesicle, is greater than the free energy of the bilayer, Fbilayer. These free energies are defined as:24 *

:

Effects of Flow and Bulk Vesicle Concentration on Supported Lipid

Recommend Documents